[PATCH 0/2] sound: soc: replace wq users and add WQ_PERCPU to alloc_workqueue() users

[PATCH 0/2] sound: soc: replace wq users and add WQ_PERCPU to alloc_workqueue() users

[Marco Crivellari]

Hi!

Below is a summary of a discussion about the Workqueue API and cpu isolation
considerations. Details and more information are available here:

        "workqueue: Always use wq_select_unbound_cpu() for WORK_CPU_UNBOUND."
        https://lore.kernel.org/all/20250221112003.1dSuoGyc@linutronix.de/

=== Current situation: problems ===

Let's consider a nohz_full system with isolated CPUs: wq_unbound_cpumask is
set to the housekeeping CPUs, for !WQ_UNBOUND the local CPU is selected.

This leads to different scenarios if a work item is scheduled on an isolated
CPU where "delay" value is 0 or greater then 0:
        schedule_delayed_work(, 0);

This will be handled by __queue_work() that will queue the work item on the
current local (isolated) CPU, while:

        schedule_delayed_work(, 1);

Will move the timer on an housekeeping CPU, and schedule the work there.

Currently if a user enqueue a work item using schedule_delayed_work() the
used wq is "system_wq" (per-cpu wq) while queue_delayed_work() use
WORK_CPU_UNBOUND (used when a cpu is not specified). The same applies to
schedule_work() that is using system_wq and queue_work(), that makes use
again of WORK_CPU_UNBOUND.

This lack of consistentcy cannot be addressed without refactoring the API.

=== Plan and future plans ===

This patchset is the first stone on a refactoring needed in order to
address the points aforementioned; it will have a positive impact also
on the cpu isolation, in the long term, moving away percpu workqueue in
favor to an unbound model.

These are the main steps:
1)  API refactoring (that this patch is introducing)
    -   Make more clear and uniform the system wq names, both per-cpu and
        unbound. This to avoid any possible confusion on what should be
        used.

    -   Introduction of WQ_PERCPU: this flag is the complement of WQ_UNBOUND,
        introduced in this patchset and used on all the callers that are not
        currently using WQ_UNBOUND.

        WQ_UNBOUND will be removed in a future release cycle.

        Most users don't need to be per-cpu, because they don't have
        locality requirements, because of that, a next future step will be
        make "unbound" the default behavior.

2)  Check who really needs to be per-cpu
    -   Remove the WQ_PERCPU flag when is not strictly required.

3)  Add a new API (prefer local cpu)
    -   There are users that don't require a local execution, like mentioned
        above; despite that, local execution yeld to performance gain.

        This new API will prefer the local execution, without requiring it.

=== Introduced Changes by this series ===

1) [P 1-2] Replace use of system_wq and system_unbound_wq

        system_wq is a per-CPU workqueue, but his name is not clear.
        system_unbound_wq is to be used when locality is not required.

        Because of that, system_wq has been renamed in system_percpu_wq, and
        system_unbound_wq has been renamed in system_dfl_wq.

=== For Maintainers ===

There are prerequisites for this series, already merged in the master branch.
The commits are:

128ea9f6ccfb6960293ae4212f4f97165e42222d ("workqueue: Add system_percpu_wq and
system_dfl_wq")

930c2ea566aff59e962c50b2421d5fcc3b98b8be ("workqueue: Add new WQ_PERCPU flag")


Thanks!

Marco Crivellari (2):
  ASoC: replace use of system_unbound_wq with system_dfl_wq
  ASoC: replace use of system_wq with system_percpu_wq

 sound/soc/codecs/aw88081.c          |  2 +-
 sound/soc/codecs/aw88166.c          |  2 +-
 sound/soc/codecs/aw88261.c          |  2 +-
 sound/soc/codecs/aw88395/aw88395.c  |  2 +-
 sound/soc/codecs/aw88399.c          |  2 +-
 sound/soc/codecs/cs42l43-jack.c     |  6 +++---
 sound/soc/codecs/cs42l43.c          |  4 ++--
 sound/soc/codecs/es8326.c           | 12 ++++++------
 sound/soc/codecs/rt5663.c           |  6 +++---
 sound/soc/codecs/wm_adsp.c          |  2 +-
 sound/soc/intel/boards/sof_es8336.c |  2 +-
 sound/soc/sof/intel/cnl.c           |  2 +-
 sound/soc/sof/intel/hda-ipc.c       |  2 +-
 13 files changed, 23 insertions(+), 23 deletions(-)

-- 
2.51.0

[PATCH 1/2] ASoC: replace use of system_unbound_wq with system_dfl_wq

[Marco Crivellari]

Currently if a user enqueue a work item using schedule_delayed_work() the
used wq is "system_wq" (per-cpu wq) while queue_delayed_work() use
WORK_CPU_UNBOUND (used when a cpu is not specified). The same applies to
schedule_work() that is using system_wq and queue_work(), that makes use
again of WORK_CPU_UNBOUND.

This lack of consistentcy cannot be addressed without refactoring the API.

system_unbound_wq should be the default workqueue so as not to enforce
locality constraints for random work whenever it's not required.

Adding system_dfl_wq to encourage its use when unbound work should be used.

queue_work() / queue_delayed_work() / mod_delayed_work() will now use the
new unbound wq: whether the user still use the old wq a warn will be
printed along with a wq redirect to the new one.

The old system_unbound_wq will be kept for a few release cycles.

Suggested-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Marco Crivellari <marco.crivellari@suse.com>
---
 sound/soc/codecs/wm_adsp.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sound/soc/codecs/wm_adsp.c b/sound/soc/codecs/wm_adsp.c
index 91c8697c29c3..c8fff8496ede 100644
--- a/sound/soc/codecs/wm_adsp.c
+++ b/sound/soc/codecs/wm_adsp.c
@@ -1044,7 +1044,7 @@ int wm_adsp_early_event(struct snd_soc_dapm_widget *w,
 
 	switch (event) {
 	case SND_SOC_DAPM_PRE_PMU:
-		queue_work(system_unbound_wq, &dsp->boot_work);
+		queue_work(system_dfl_wq, &dsp->boot_work);
 		break;
 	case SND_SOC_DAPM_PRE_PMD:
 		wm_adsp_power_down(dsp);
-- 
2.51.0

[PATCH 2/2] ASoC: replace use of system_wq with system_percpu_wq

[Marco Crivellari]

Currently if a user enqueue a work item using schedule_delayed_work() the
used wq is "system_wq" (per-cpu wq) while queue_delayed_work() use
WORK_CPU_UNBOUND (used when a cpu is not specified). The same applies to
schedule_work() that is using system_wq and queue_work(), that makes use
again of WORK_CPU_UNBOUND.

This lack of consistentcy cannot be addressed without refactoring the API.

system_wq is a per-CPU worqueue, yet nothing in its name tells about that
CPU affinity constraint, which is very often not required by users. Make
it clear by adding a system_percpu_wq.

queue_work() / queue_delayed_work() mod_delayed_work() will now use the
new per-cpu wq: whether the user still stick on the old name a warn will
be printed along a wq redirect to the new one.

This patch add the new system_percpu_wq except for mm, fs and net
subsystem, whom are handled in separated patches.

The old wq will be kept for a few release cylces.

Suggested-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Marco Crivellari <marco.crivellari@suse.com>
---
 sound/soc/codecs/aw88081.c          |  2 +-
 sound/soc/codecs/aw88166.c          |  2 +-
 sound/soc/codecs/aw88261.c          |  2 +-
 sound/soc/codecs/aw88395/aw88395.c  |  2 +-
 sound/soc/codecs/aw88399.c          |  2 +-
 sound/soc/codecs/cs42l43-jack.c     |  6 +++---
 sound/soc/codecs/cs42l43.c          |  4 ++--
 sound/soc/codecs/es8326.c           | 12 ++++++------
 sound/soc/codecs/rt5663.c           |  6 +++---
 sound/soc/intel/boards/sof_es8336.c |  2 +-
 sound/soc/sof/intel/cnl.c           |  2 +-
 sound/soc/sof/intel/hda-ipc.c       |  2 +-
 12 files changed, 22 insertions(+), 22 deletions(-)

diff --git a/sound/soc/codecs/aw88081.c b/sound/soc/codecs/aw88081.c
index ad16ab6812cd..e61c58dcd606 100644
--- a/sound/soc/codecs/aw88081.c
+++ b/sound/soc/codecs/aw88081.c
@@ -779,7 +779,7 @@ static void aw88081_start(struct aw88081 *aw88081, bool sync_start)
 	if (sync_start == AW88081_SYNC_START)
 		aw88081_start_pa(aw88081);
 	else
-		queue_delayed_work(system_wq,
+		queue_delayed_work(system_percpu_wq,
 			&aw88081->start_work,
 			AW88081_START_WORK_DELAY_MS);
 }
diff --git a/sound/soc/codecs/aw88166.c b/sound/soc/codecs/aw88166.c
index 6c50c4a18b6a..c9c3ebb9a739 100644
--- a/sound/soc/codecs/aw88166.c
+++ b/sound/soc/codecs/aw88166.c
@@ -1313,7 +1313,7 @@ static void aw88166_start(struct aw88166 *aw88166, bool sync_start)
 	if (sync_start == AW88166_SYNC_START)
 		aw88166_start_pa(aw88166);
 	else
-		queue_delayed_work(system_wq,
+		queue_delayed_work(system_percpu_wq,
 			&aw88166->start_work,
 			AW88166_START_WORK_DELAY_MS);
 }
diff --git a/sound/soc/codecs/aw88261.c b/sound/soc/codecs/aw88261.c
index fb99871578c5..c8e62af8949e 100644
--- a/sound/soc/codecs/aw88261.c
+++ b/sound/soc/codecs/aw88261.c
@@ -705,7 +705,7 @@ static void aw88261_start(struct aw88261 *aw88261, bool sync_start)
 	if (sync_start == AW88261_SYNC_START)
 		aw88261_start_pa(aw88261);
 	else
-		queue_delayed_work(system_wq,
+		queue_delayed_work(system_percpu_wq,
 			&aw88261->start_work,
 			AW88261_START_WORK_DELAY_MS);
 }
diff --git a/sound/soc/codecs/aw88395/aw88395.c b/sound/soc/codecs/aw88395/aw88395.c
index aea44a199b98..c6fe69cc5e73 100644
--- a/sound/soc/codecs/aw88395/aw88395.c
+++ b/sound/soc/codecs/aw88395/aw88395.c
@@ -75,7 +75,7 @@ static void aw88395_start(struct aw88395 *aw88395, bool sync_start)
 	if (sync_start == AW88395_SYNC_START)
 		aw88395_start_pa(aw88395);
 	else
-		queue_delayed_work(system_wq,
+		queue_delayed_work(system_percpu_wq,
 			&aw88395->start_work,
 			AW88395_START_WORK_DELAY_MS);
 }
diff --git a/sound/soc/codecs/aw88399.c b/sound/soc/codecs/aw88399.c
index ee3cc2a95f85..dfa8ce355e3c 100644
--- a/sound/soc/codecs/aw88399.c
+++ b/sound/soc/codecs/aw88399.c
@@ -1281,7 +1281,7 @@ static void aw88399_start(struct aw88399 *aw88399, bool sync_start)
 	if (sync_start == AW88399_SYNC_START)
 		aw88399_start_pa(aw88399);
 	else
-		queue_delayed_work(system_wq,
+		queue_delayed_work(system_percpu_wq,
 			&aw88399->start_work,
 			AW88399_START_WORK_DELAY_MS);
 }
diff --git a/sound/soc/codecs/cs42l43-jack.c b/sound/soc/codecs/cs42l43-jack.c
index ac19a572fe70..38c73c8dcc45 100644
--- a/sound/soc/codecs/cs42l43-jack.c
+++ b/sound/soc/codecs/cs42l43-jack.c
@@ -301,7 +301,7 @@ irqreturn_t cs42l43_bias_detect_clamp(int irq, void *data)
 {
 	struct cs42l43_codec *priv = data;
 
-	queue_delayed_work(system_wq, &priv->bias_sense_timeout,
+	queue_delayed_work(system_percpu_wq, &priv->bias_sense_timeout,
 			   msecs_to_jiffies(1000));
 
 	return IRQ_HANDLED;
@@ -432,7 +432,7 @@ irqreturn_t cs42l43_button_press(int irq, void *data)
 	struct cs42l43_codec *priv = data;
 
 	// Wait for 2 full cycles of comb filter to ensure good reading
-	queue_delayed_work(system_wq, &priv->button_press_work,
+	queue_delayed_work(system_percpu_wq, &priv->button_press_work,
 			   msecs_to_jiffies(20));
 
 	return IRQ_HANDLED;
@@ -470,7 +470,7 @@ irqreturn_t cs42l43_button_release(int irq, void *data)
 {
 	struct cs42l43_codec *priv = data;
 
-	queue_work(system_wq, &priv->button_release_work);
+	queue_work(system_percpu_wq, &priv->button_release_work);
 
 	return IRQ_HANDLED;
 }
diff --git a/sound/soc/codecs/cs42l43.c b/sound/soc/codecs/cs42l43.c
index ea84ac64c775..105ad53bae0c 100644
--- a/sound/soc/codecs/cs42l43.c
+++ b/sound/soc/codecs/cs42l43.c
@@ -161,7 +161,7 @@ static void cs42l43_hp_ilimit_clear_work(struct work_struct *work)
 	priv->hp_ilimit_count--;
 
 	if (priv->hp_ilimit_count)
-		queue_delayed_work(system_wq, &priv->hp_ilimit_clear_work,
+		queue_delayed_work(system_percpu_wq, &priv->hp_ilimit_clear_work,
 				   msecs_to_jiffies(CS42L43_HP_ILIMIT_DECAY_MS));
 
 	snd_soc_dapm_mutex_unlock(dapm);
@@ -178,7 +178,7 @@ static void cs42l43_hp_ilimit_work(struct work_struct *work)
 
 	if (priv->hp_ilimit_count < CS42L43_HP_ILIMIT_MAX_COUNT) {
 		if (!priv->hp_ilimit_count)
-			queue_delayed_work(system_wq, &priv->hp_ilimit_clear_work,
+			queue_delayed_work(system_percpu_wq, &priv->hp_ilimit_clear_work,
 					   msecs_to_jiffies(CS42L43_HP_ILIMIT_DECAY_MS));
 
 		priv->hp_ilimit_count++;
diff --git a/sound/soc/codecs/es8326.c b/sound/soc/codecs/es8326.c
index 066d92b54312..4ba4de184d2c 100644
--- a/sound/soc/codecs/es8326.c
+++ b/sound/soc/codecs/es8326.c
@@ -812,12 +812,12 @@ static void es8326_jack_button_handler(struct work_struct *work)
 			press_count = 0;
 		}
 		button_to_report = cur_button;
-		queue_delayed_work(system_wq, &es8326->button_press_work,
+		queue_delayed_work(system_percpu_wq, &es8326->button_press_work,
 				   msecs_to_jiffies(35));
 	} else if (prev_button != cur_button) {
 		/* mismatch, detect again */
 		prev_button = cur_button;
-		queue_delayed_work(system_wq, &es8326->button_press_work,
+		queue_delayed_work(system_percpu_wq, &es8326->button_press_work,
 				   msecs_to_jiffies(35));
 	} else {
 		/* released or no pressed */
@@ -912,7 +912,7 @@ static void es8326_jack_detect_handler(struct work_struct *work)
 					(ES8326_INT_SRC_PIN9 | ES8326_INT_SRC_BUTTON));
 			regmap_write(es8326->regmap, ES8326_SYS_BIAS, 0x1f);
 			regmap_update_bits(es8326->regmap, ES8326_HP_DRIVER_REF, 0x0f, 0x0d);
-			queue_delayed_work(system_wq, &es8326->jack_detect_work,
+			queue_delayed_work(system_percpu_wq, &es8326->jack_detect_work,
 					msecs_to_jiffies(400));
 			es8326->hp = 1;
 			goto exit;
@@ -923,7 +923,7 @@ static void es8326_jack_detect_handler(struct work_struct *work)
 			regmap_write(es8326->regmap, ES8326_INT_SOURCE,
 					(ES8326_INT_SRC_PIN9 | ES8326_INT_SRC_BUTTON));
 			es8326_enable_micbias(es8326->component);
-			queue_delayed_work(system_wq, &es8326->button_press_work, 10);
+			queue_delayed_work(system_percpu_wq, &es8326->button_press_work, 10);
 			goto exit;
 		}
 		if ((iface & ES8326_HPBUTTON_FLAG) == 0x01) {
@@ -958,10 +958,10 @@ static irqreturn_t es8326_irq(int irq, void *dev_id)
 		goto out;
 
 	if (es8326->jack->status & SND_JACK_HEADSET)
-		queue_delayed_work(system_wq, &es8326->jack_detect_work,
+		queue_delayed_work(system_percpu_wq, &es8326->jack_detect_work,
 				   msecs_to_jiffies(10));
 	else
-		queue_delayed_work(system_wq, &es8326->jack_detect_work,
+		queue_delayed_work(system_percpu_wq, &es8326->jack_detect_work,
 				   msecs_to_jiffies(300));
 
 out:
diff --git a/sound/soc/codecs/rt5663.c b/sound/soc/codecs/rt5663.c
index 45057562c0c8..44cfec76ad96 100644
--- a/sound/soc/codecs/rt5663.c
+++ b/sound/soc/codecs/rt5663.c
@@ -1859,7 +1859,7 @@ static irqreturn_t rt5663_irq(int irq, void *data)
 	dev_dbg(regmap_get_device(rt5663->regmap), "%s IRQ queue work\n",
 		__func__);
 
-	queue_delayed_work(system_wq, &rt5663->jack_detect_work,
+	queue_delayed_work(system_percpu_wq, &rt5663->jack_detect_work,
 		msecs_to_jiffies(250));
 
 	return IRQ_HANDLED;
@@ -1974,7 +1974,7 @@ static void rt5663_jack_detect_work(struct work_struct *work)
 				cancel_delayed_work_sync(
 					&rt5663->jd_unplug_work);
 			} else {
-				queue_delayed_work(system_wq,
+				queue_delayed_work(system_percpu_wq,
 					&rt5663->jd_unplug_work,
 					msecs_to_jiffies(500));
 			}
@@ -2024,7 +2024,7 @@ static void rt5663_jd_unplug_work(struct work_struct *work)
 				    SND_JACK_BTN_0 | SND_JACK_BTN_1 |
 				    SND_JACK_BTN_2 | SND_JACK_BTN_3);
 	} else {
-		queue_delayed_work(system_wq, &rt5663->jd_unplug_work,
+		queue_delayed_work(system_percpu_wq, &rt5663->jd_unplug_work,
 			msecs_to_jiffies(500));
 	}
 }
diff --git a/sound/soc/intel/boards/sof_es8336.c b/sound/soc/intel/boards/sof_es8336.c
index a0b3679b17b4..e60dd85f5552 100644
--- a/sound/soc/intel/boards/sof_es8336.c
+++ b/sound/soc/intel/boards/sof_es8336.c
@@ -163,7 +163,7 @@ static int sof_es8316_speaker_power_event(struct snd_soc_dapm_widget *w,
 
 	priv->speaker_en = !SND_SOC_DAPM_EVENT_ON(event);
 
-	queue_delayed_work(system_wq, &priv->pcm_pop_work, msecs_to_jiffies(70));
+	queue_delayed_work(system_percpu_wq, &priv->pcm_pop_work, msecs_to_jiffies(70));
 	return 0;
 }
 
diff --git a/sound/soc/sof/intel/cnl.c b/sound/soc/sof/intel/cnl.c
index 385e5339f0a4..207eb18560dd 100644
--- a/sound/soc/sof/intel/cnl.c
+++ b/sound/soc/sof/intel/cnl.c
@@ -329,7 +329,7 @@ int cnl_ipc_send_msg(struct snd_sof_dev *sdev, struct snd_sof_ipc_msg *msg)
 	 * CTX_SAVE IPC, which is sent before the DSP enters D3.
 	 */
 	if (hdr->cmd != (SOF_IPC_GLB_PM_MSG | SOF_IPC_PM_CTX_SAVE))
-		mod_delayed_work(system_wq, &hdev->d0i3_work,
+		mod_delayed_work(system_percpu_wq, &hdev->d0i3_work,
 				 msecs_to_jiffies(SOF_HDA_D0I3_WORK_DELAY_MS));
 
 	return 0;
diff --git a/sound/soc/sof/intel/hda-ipc.c b/sound/soc/sof/intel/hda-ipc.c
index f3fbf43a70c2..d8fde18145b4 100644
--- a/sound/soc/sof/intel/hda-ipc.c
+++ b/sound/soc/sof/intel/hda-ipc.c
@@ -96,7 +96,7 @@ void hda_dsp_ipc4_schedule_d0i3_work(struct sof_intel_hda_dev *hdev,
 	if (hda_dsp_ipc4_pm_msg(msg_data->primary))
 		return;
 
-	mod_delayed_work(system_wq, &hdev->d0i3_work,
+	mod_delayed_work(system_percpu_wq, &hdev->d0i3_work,
 			 msecs_to_jiffies(SOF_HDA_D0I3_WORK_DELAY_MS));
 }
 EXPORT_SYMBOL_NS(hda_dsp_ipc4_schedule_d0i3_work, "SND_SOC_SOF_INTEL_HDA_COMMON");
-- 
2.51.0

Re: [PATCH 2/2] ASoC: replace use of system_wq with system_percpu_wq

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 539 bytes --]

On Fri, Sep 05, 2025 at 11:10:16AM +0200, Marco Crivellari wrote:

> +++ b/sound/soc/codecs/aw88081.c
> @@ -779,7 +779,7 @@ static void aw88081_start(struct aw88081 *aw88081, bool sync_start)
>  	if (sync_start == AW88081_SYNC_START)
>  		aw88081_start_pa(aw88081);
>  	else
> -		queue_delayed_work(system_wq,
> +		queue_delayed_work(system_percpu_wq,
>  			&aw88081->start_work,
>  			AW88081_START_WORK_DELAY_MS);
>  }

I can't tell why you're putting all these onto the percpu workqeue, I
would be astonished if any of the users cared.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: (subset) [PATCH 0/2] sound: soc: replace wq users and add WQ_PERCPU to alloc_workqueue() users

[Mark Brown]

On Fri, 05 Sep 2025 11:10:14 +0200, Marco Crivellari wrote:
> Below is a summary of a discussion about the Workqueue API and cpu isolation
> considerations. Details and more information are available here:
> 
>         "workqueue: Always use wq_select_unbound_cpu() for WORK_CPU_UNBOUND."
>         https://lore.kernel.org/all/20250221112003.1dSuoGyc@linutronix.de/
> 
> === Current situation: problems ===
> 
> [...]

Applied to

   https://git.kernel.org/pub/scm/linux/kernel/git/broonie/sound.git for-next

Thanks!

[1/2] ASoC: replace use of system_unbound_wq with system_dfl_wq
      commit: 9e5eb8b49ffe3c173bf7b8c338a57dfa09fb4634

All being well this means that it will be integrated into the linux-next
tree (usually sometime in the next 24 hours) and sent to Linus during
the next merge window (or sooner if it is a bug fix), however if
problems are discovered then the patch may be dropped or reverted.

You may get further e-mails resulting from automated or manual testing
and review of the tree, please engage with people reporting problems and
send followup patches addressing any issues that are reported if needed.

If any updates are required or you are submitting further changes they
should be sent as incremental updates against current git, existing
patches will not be replaced.

Please add any relevant lists and maintainers to the CCs when replying
to this mail.

Thanks,
Mark

Re: [PATCH 2/2] ASoC: replace use of system_wq with system_percpu_wq

[Marco Crivellari]

Hello Mark,

The above change is just a 1:1 conversion.
system_wq is a per-cpu wq, so we only converted the name with the new one.
system_wq is still present in the code, for a few release cycles.

If the users don't care to be per-cpu, system_dfl_wq (the new unbound
wq) can be used instead.

This is just the first step of the conversion.

Thanks!


On Fri, Sep 5, 2025 at 12:54 PM Mark Brown <broonie@kernel.org> wrote:
>
> On Fri, Sep 05, 2025 at 11:10:16AM +0200, Marco Crivellari wrote:
>
> > +++ b/sound/soc/codecs/aw88081.c
> > @@ -779,7 +779,7 @@ static void aw88081_start(struct aw88081 *aw88081, bool sync_start)
> >       if (sync_start == AW88081_SYNC_START)
> >               aw88081_start_pa(aw88081);
> >       else
> > -             queue_delayed_work(system_wq,
> > +             queue_delayed_work(system_percpu_wq,
> >                       &aw88081->start_work,
> >                       AW88081_START_WORK_DELAY_MS);
> >  }
>
> I can't tell why you're putting all these onto the percpu workqeue, I
> would be astonished if any of the users cared.



-- 

Marco Crivellari

L3 Support Engineer, Technology & Product




marco.crivellari@suse.com

Re: [PATCH 2/2] ASoC: replace use of system_wq with system_percpu_wq

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 893 bytes --]

On Fri, Sep 05, 2025 at 06:00:04PM +0200, Marco Crivellari wrote:

> Hello Mark,

Please don't top post, reply in line with needed context.  This allows
readers to readily follow the flow of conversation and understand what
you are talking about and also helps ensure that everything in the
discussion is being addressed.

> The above change is just a 1:1 conversion.
> system_wq is a per-cpu wq, so we only converted the name with the new one.
> system_wq is still present in the code, for a few release cycles.

> If the users don't care to be per-cpu, system_dfl_wq (the new unbound
> wq) can be used instead.

> This is just the first step of the conversion.

If we're going to convert everything I'd rather we do a sensible
conversion now than have multiple rounds of conversions.  I strongly
suspect that most if not all of the current users don't know or care
that system_wq is percpu.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: [PATCH 2/2] ASoC: replace use of system_wq with system_percpu_wq

[Marco Crivellari]

On Fri, Sep 5, 2025 at 6:05 PM Mark Brown <broonie@kernel.org> wrote:
>
> On Fri, Sep 05, 2025 at 06:00:04PM +0200, Marco Crivellari wrote:
>
> > Hello Mark,
>
> Please don't top post, reply in line with needed context.  This allows
> readers to readily follow the flow of conversation and understand what
> you are talking about and also helps ensure that everything in the
> discussion is being addressed.

Hello Mark,

Noted, sorry for that.

> If we're going to convert everything I'd rather we do a sensible
> conversion now than have multiple rounds of conversions.  I strongly
> suspect that most if not all of the current users don't know or care
> that system_wq is percpu.

Sure, we can do so. I will send the v2 hopefully with 1 patch only
(meaning that every system_wq can be converted to system_dfl_wq).

Thank you!

--

Marco Crivellari

L3 Support Engineer, Technology & Product

marco.crivellari@suse.com

Re: [PATCH 2/2] ASoC: replace use of system_wq with system_percpu_wq

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 557 bytes --]

On Mon, Sep 08, 2025 at 04:52:53PM +0200, Marco Crivellari wrote:
> On Fri, Sep 5, 2025 at 6:05 PM Mark Brown <broonie@kernel.org> wrote:

> > If we're going to convert everything I'd rather we do a sensible
> > conversion now than have multiple rounds of conversions.  I strongly
> > suspect that most if not all of the current users don't know or care
> > that system_wq is percpu.

> Sure, we can do so. I will send the v2 hopefully with 1 patch only
> (meaning that every system_wq can be converted to system_dfl_wq).

Sounds good thanks!

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]